A number of recent conversations have reminded me that the links between decision analysis and quality control are not generally well understood. The most basic link is probability, the mathematical theory pioneered in the sixteenth-century by Gerolamo Cardano in his attempts to analyze games of chance and by later philosophers and mathematicians attempting to explain randomness.
Randomness takes us to the next important link between decision evaluation and quality improvement: Monte Carlo simulation, a computational device to bring randomness into predictions about real world events. Because it uses repeated computation and random numbers Monte Carlo simulation came into widespread use only with the advent of the desktop computer.
Decision analysis uses Monte Carlo software to predict the future outcomes of a decision made under uncertainty. It answers what-if questions with probabilities and thus allows a decision maker to determine the best probable course of action. Risk simulation, the same method by a different name, quantifies the “risk” in any defined scenario in terms of probability.
And it is the term scenario that is our last connection to quality improvement. The field of quality improvement is concerned with perfecting processes, and any process can be viewed as a scenario. Each event in a scenario requires an answer to a what-if problem, and these events can be strung together to predict the probable results of a process. In other words, a process is a series of decisions made under uncertainty, and simulation of a particular process–say, reserve estimation exploration for oil–is essentially a series of decision analyses. Process–or quality–improvement results from the putting the probabilities in these analyses to use.
